A-levels are not an accurate test of a student's abilities and potential, academics have said.

Students would need to be tested for 40 hours for each subject for universities to have a true picture of their abilities, according to academics from King's College, London.

The release of their findings coincides with a renewed debate about exam results and whether standards are slipping.

Professor Dylan William and Professor Paul Black believe standards have been "broadly maintained".

But they say universities need other evidence of a student's abilities besides A-level grades.

Margaret Hodge agrees A-levels are not a precise indicator of talent

Professor William says grades might be improving because students are working harder and schools are increasingly "teaching to the test".

He said: "Whether standards are going up or down is not the issue - standards have been broadly maintained.

"The problem is we just don't know how accurate examination grades are for individual students, and they are of only limited use as predictors of future performance."

The academics said research in the 1970s showed A-levels were only accurate to plus or minus a grade.

They say the only way to make A-levels accurate to a 10th of a grade would be to increase the amount of exams students had to take to about 40 hours for each subject.

Prof William said this would not be popular with students so instead people would have to start trusting the judgements of teachers and university admissions tutors.

Some universities, including Kings College, use aptitude tests to gauge potential.

They are routinely used at American universities.

Professor William said although candidates for medicine at Kings usually have to get two As and a B, the college had accepted nine students from state schools who had Cs and Ds but scored well in tests on science reasoning.

Philanthropist

This is an approach advocated by the Higher Education Minster Margaret Hodge, who has said universities should consider lowering their A-level requirements for students from working-class backgrounds.

In a speech in April, she said A-levels were poor at measuring a student's potential.

At Bristol University, admissions tutors look at the average A-level scores of children at a candidate's school before awarding a place.

A pupil who had done much better than their school friends would be given a place even if their A-level results did not match those of other people previously accepted by Bristol.

The use of aptitude tests to spot talented students is advocated by Peter Lampl, the philanthropist behind the Sutton Trust educational charity.

He says more universities should use them to open their doors to bright children from low-income families.